How to OCR a PDF: Extract Text from Scanned Documents with C# .NET OCR PDF

How Can I Perform OCR on a PDF in C#?

First, install the IronOCR library via NuGet Package Manager to add this powerful OCR engine to your system:

Install-Package IronOcr

The following example demonstrates how to load a PDF file and recognize text from an entire scanned document:

using IronOcr;
// Initialize the OCR engine
IronTesseract ocr = new IronTesseract();
// Load the PDF and perform OCR
using var pdfInput = new OcrPdfInput("scanned-report.pdf");
OcrResult result = ocr.Read(pdfInput);
// Output the extracted text
string extractedText = result.Text;
Console.WriteLine(extractedText);

using IronOcr;
// Initialize the OCR engine
IronTesseract ocr = new IronTesseract();
// Load the PDF and perform OCR
using var pdfInput = new OcrPdfInput("scanned-report.pdf");
OcrResult result = ocr.Read(pdfInput);
// Output the extracted text
string extractedText = result.Text;
Console.WriteLine(extractedText);

IRON VB CONVERTER ERROR developers@ironsoftware.com

The IronTesseract class serves as the primary OCR engine, wrapping Tesseract 5 with optimizations for .NET Core and .NET Framework applications. The OcrPdfInput object handles PDF loading and page rendering internally, eliminating the need to convert image formats manually. When you call the Read method, the OCR process analyzes each page and returns an OcrResult containing the extracted text as a string, along with structured data about paragraphs, lines, words, and their positions. Users can then save the output to a TXT file, target folder, Word documents, or use the API to process the data further.

How Do I Read Specific Pages from a PDF?

Processing large text documents becomes more efficient when you target only the pages you need. Pass a list of page indices to the PageIndices parameter to convert scanned PDF pages selectively:

using IronOcr;
using System.Collections.Generic;
IronTesseract ocr = new IronTesseract();
// Specify pages to process (zero-based indexing)
List<int> targetPages = new List<int>() { 0, 2, 4 };
using var pdfInput = new OcrPdfInput("lengthy-document.pdf", PageIndices: targetPages);
OcrResult result = ocr.Read(pdfInput);
// Save or process the OCR results
Console.WriteLine(result.Text);

using IronOcr;
using System.Collections.Generic;
IronTesseract ocr = new IronTesseract();
// Specify pages to process (zero-based indexing)
List<int> targetPages = new List<int>() { 0, 2, 4 };
using var pdfInput = new OcrPdfInput("lengthy-document.pdf", PageIndices: targetPages);
OcrResult result = ocr.Read(pdfInput);
// Save or process the OCR results
Console.WriteLine(result.Text);

IRON VB CONVERTER ERROR developers@ironsoftware.com

Note that IronOCR uses zero-based indexing, so page 0 represents the first page of your PDF document. This selective approach reduces processing time and memory consumption when working with multi-page scanned documents where only specific sections contain relevant searchable data.

How Can I Extract Data from a Specific Region?

Invoice processing, form digitization, and document parsing often require extracting text from defined areas rather than entire pages. This OCR tool allows you to create targeted scans using the ContentAreas parameter, which accepts an array of rectangles specifying the regions to process:

using IronOcr;
using IronSoftware.Drawing;
using System;
IronTesseract ocr = new IronTesseract();
// Define the scan region (x, y, width, height in pixels)
Rectangle[] invoiceFields = {
    new Rectangle(130, 290, 250, 50)   // Invoice number area
};
using var pdfInput = new OcrPdfInput("invoice.pdf", ContentAreas: invoiceFields);
OcrResult result = ocr.Read(pdfInput);
// Extract and output the structured data
Console.WriteLine(result.Text);

using IronOcr;
using IronSoftware.Drawing;
using System;
IronTesseract ocr = new IronTesseract();
// Define the scan region (x, y, width, height in pixels)
Rectangle[] invoiceFields = {
    new Rectangle(130, 290, 250, 50)   // Invoice number area
};
using var pdfInput = new OcrPdfInput("invoice.pdf", ContentAreas: invoiceFields);
OcrResult result = ocr.Read(pdfInput);
// Extract and output the structured data
Console.WriteLine(result.Text);

IRON VB CONVERTER ERROR developers@ironsoftware.com

The Rectangle constructor accepts four parameters: X position, Y position, width, and height—all measured in pixels from the top-left corner of the page. This targeted text recognition approach dramatically improves both speed and accuracy by focusing the OCR engine on specific content areas rather than processing irrelevant background elements. For batch invoice processing, combine region extraction with iterating through result pages to build editable structured data from multiple PDF files.

How Do I Improve OCR Accuracy on Scanned Documents?

Real-world scanned paper documents often arrive with quality issues: skewed pages, low resolution, or digital noise from the scanning software. IronOCR includes preprocessing filters that address these challenges and help convert image quality problems into accurate text conversion:

using IronOcr;
IronTesseract ocr = new IronTesseract();
using var input = new OcrInput();
// Load PDF with higher DPI for better text recognition
input.LoadPdf("poor-quality-scan.pdf", DPI: 300);
// Apply image correction filters to process scanned images
input.Deskew();   // Straighten rotated pages
input.DeNoise();  // Remove scanning artifacts
OcrResult result = ocr.Read(input);
Console.WriteLine(result.Text);

using IronOcr;
IronTesseract ocr = new IronTesseract();
using var input = new OcrInput();
// Load PDF with higher DPI for better text recognition
input.LoadPdf("poor-quality-scan.pdf", DPI: 300);
// Apply image correction filters to process scanned images
input.Deskew();   // Straighten rotated pages
input.DeNoise();  // Remove scanning artifacts
OcrResult result = ocr.Read(input);
Console.WriteLine(result.Text);

IRON VB CONVERTER ERROR developers@ironsoftware.com

The DPI parameter controls the resolution at which PDF pages are rendered before the OCR process—higher values (200-300 DPI) improve accuracy for documents with small text. The Deskew method automatically detects and corrects page rotation, while DeNoise removes speckles and artifacts that interfere with optical character recognition (OCR). For documents requiring additional image adjustments, IronOCR provides contrast enhancement, binarization, and other tools to edit image quality.

This .NET OCR library also handles password-protected PDF documents by accepting credentials during input construction. The software supports 125+ language packs, enabling OCR on international documents. Beyond standard PDF files, IronOCR can process PNG, TIFF (including multipage TIFF), and other image format files. Deployment works seamlessly on Windows, Linux, macOS, and cloud platforms including Azure and Docker containers.

Conclusion

IronOCR transforms the complex task of PDF text extraction into a straightforward operation. From basic document reading to targeted region extraction and preprocessing for challenging scanned images, this OCR library handles the technical complexity while exposing a clean C# API that works across .NET Core and .NET Framework.

The code examples above demonstrate core functionality, but IronOCR extends further with barcode and QR code reading, searchable PDF creation to convert scanned PDF files into editable searchable documents, and structured data output including confidence scores and text positioning. Explore the complete API reference for advanced implementations—or try the free pro version features during your trial.

Purchase a license to deploy IronOCR in .NET applications production environments, or chat with our engineering team for project-specific guidance.

Ready to perform OCR in your .NET applications? Start with a free trial to explore the full feature set and download the SDK.